[dashboard] GPU limits are not taken into account

[wbuchwalter]

In the dashboard if I specify a number of GPU for my master/worker, this is not taken into account at the creation and pods will have no resource limits.

[wbuchwalter]

This needs to happen first: #450

[ScorpioCPH]

Hi, @wbuchwalter Do you think we need a separated repo for dashboard, such as kubeflow/tf-dashboard.

@jlewi WDYT?

[gaocegege]

@ScorpioCPH We have some discussions in #389 . I think we could create a repo for the dashboard.

[ScorpioCPH]

@gaocegege Good!

[jlewi]

@wbuchwalter Should this be P2 or P1? P1 means it will get done by our 0.2 release which is end of June.

[jlewi]

/label priority/p2

[jlewi]

Removing this from the 0.2 release because it doesn't look like its going to make the cut at this point.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.